A note on the magnitude of the feedback effect in environmentally extended multi-region input-output tables

Moran, D., Wood, R. and Rodrigues, J. F. D. (2018). Journal of Industrial Ecology, vol. 22, no. 3: 532–39. DOI:10.1111/jiec.12658

Would feedback loops compromise environmental indicators generated from the PRINCE hybrid model?


Global multiregional input-output (MRIO) tables have been developed to capture international spillover effects due to demand in one country and production in other countries. International spillovers have been growing and have become so dominant, especially in environmental analysis, that their inclusion is essential when analyzing impacts of consumption.

MRIO tables give full coverage of the world economy, but do not always respect the official data of a given country. When international spillovers also cause increased production in the country of demand, we see what are known as “feedback effects.”

A coupled model such as that developed for PRINCE using an official foreground national input-output table (IOT) alongside an existing global MRIO (Exiobase) does not use the official foreground information when modeling international feedback loops. The question thus arises: are the feedback loops large enough to compromise the accuracy of results for different environmental impacts generated from the model?

The authors looked specifically at the amount of domestic production that is embodied in imports back into a region. They found that for emissions the feedbacks were small, usually <2% of the total import footprint, though up to 6%+ for some countries in some years for some stressors. The findings reassured them that the effect of the feedback loops on results for Sweden were were not significant, and thus using Leontief multipliers from Exiobase was an acceptable method for modeling imports into the Swedish IOT for the purposes of PRINCE.